In daily life, you often encounter various types of data related to “time”, such as:
- • A sensor records temperature every 10 seconds
- • The stock market updates prices every minute
- • A server generates hundreds of performance metrics every second
These data points share a common characteristic: they are generated in chronological order, and time is their most important attribute. This type of data is referred to as time series data, and the database specifically designed to store and analyze this type of data is called a time series database (TSDB).
In this lesson, you will learn about the basic concepts, characteristics, common application scenarios of time series databases, and how they differ from traditional databases.
What is a Time Series Database?
A time series database is a database system specifically designed to store, manage, and analyze data points arranged in chronological order. Each time series data point typically consists of the following basic components:
- • Timestamp: records the time when the data was generated
- • One or more values: such as temperature, voltage, click counts, etc.
- • Tags/Labels: used to identify the source, type, or other additional information of the data (such as device ID, location, etc.)
For example: suppose you have a device <span>device-01</span> that reports temperature and voltage data every 10 seconds, recorded as follows.
Timestamp Device ID Temperature(℃) Voltage(V)
2025-04-17 10:00:00 device-01 23.5 3.3
2025-04-17 10:00:10 device-01 23.6 3.3
2025-04-17 10:00:20 device-01 23.7 3.2This data may continuously append to the database and is often displayed on charts to show trends over time. After writing to a time series database, you can perform some query operations, such as the average temperature over the last 10 minutes, or the maximum voltage for each hour of a specific day.
Core Features of Time Series Databases
- 1. Efficient Write Performance You can quickly and continuously write a large number of data points to a time series database, which will be organized in chronological order while minimizing storage space waste.
- 2. Time is Paramount The timestamp holds the “paramount” position in a time series database. You often use queries like “data from the past hour”, “daily averages”, or “peak values between 9:00 and 10:00”.
- 3. Strong Data Compression Capability Since time series data often has continuity and similarity, time series databases can significantly reduce storage space through compression algorithms (such as differential storage, sliding windows, etc.).
- 4. Query Optimization Optimized specifically for queries based on time ranges and aggregation calculations (average, maximum, minimum, sum).
- 5. Built-in Downsampling Capability You can generate low-resolution “summaries” from the original data, such as hourly averages or daily maximums, to save query overhead.
Comparison with Traditional Databases
| Feature | Relational Database (e.g., MySQL) | Time Series Database (e.g., InfluxDB) |
| Data Organization | Organized by tables, structured | Appended by timeline, structure can vary |
| Query Method | Diverse SQL queries | More focused on time ranges and aggregation analysis |
| Performance Optimization Direction | Row/column storage, index optimization | Write performance, compression, time windows |
| Applicable Scenarios | General business data storage | High-frequency time series data analysis |
Although relational databases can also store time-related data, they fall far short of specialized time series databases in terms of write efficiency, compression ratio, and aggregation performance.
Why is TSDB Indispensable for the Internet of Things?
In Internet of Things (IoT) scenarios, the most common data is the combination of “time + state”. For example:
- • A smart wristband records heart rate every minute
- • An electric meter reports voltage and current every 10 seconds
- • A temperature and humidity sensor records indoor conditions every 5 seconds
The essence of this data is time series data, and this data typically has the following characteristics:
| Characteristic | Description |
| Continuity | Data is continuously generated, and each data point has a timestamp. |
| High Write Frequency | A large number of devices report data at high frequencies, such as every second or every minute. |
| Time-based Queries | You often need to query requests like “temperature changes in the past hour”. |
| Large Data Volume | Many devices continuously generate data, leading to significant storage pressure. |
A time series database is designed for scenarios where data is written and read in chronological order, helping you write, store, compress, and query this data more efficiently.
In actual IoT platform architecture, the time series database generally sits between the “data collection” and “visualization analysis”:
Device Collection Layer
Message Queue/MQTT
Data Processing/Rule Engine
Time Series Database (TSDB)
Visualization System (e.g., Grafana)
Alert System
You can access device data via the MQTT protocol, write it to the TSDB, and then use a visualization platform to create charts, generate reports, trigger alerts, etc.
Common Application Scenarios
Time series databases are widely used in the following fields:
| Application Field | Examples |
| 🌡️ Internet of Things (IoT) | Sensor data, smart homes, industrial equipment status. |
| 🖥️ IT Operations | CPU usage, memory consumption, network traffic monitoring. |
| 📈 Financial Markets | Stock price changes, trading volume. |
| ⚡ Energy Monitoring | Electricity usage data, photovoltaic power generation monitoring. |
| 🚀 Industrial Control | Real-time monitoring of factory production lines, predictive maintenance. |
Common Time Series Databases
Here are some mainstream time series database products:
| Name | Description | Open Source |
| InfluxDB | Lightweight and efficient, with SQL-like query syntax, highly usable. | ✅ |
| TimescaleDB | Based on PostgreSQL, combining relational model and time series features. | ✅ |
| Prometheus | Suitable for monitoring metrics collection and alerts, intersects with IoT. | ✅ |
| TDengine | Domestic TSDB, focusing on high performance for IoT. | ✅ |
| OpenTSDB | Built on HBase, suitable for large-scale data storage. | ✅ |
| QuestDB | SQL compatible, high-frequency data processing in finance/industry. | ✅ |
Conclusion
A time series database is a type of database system designed specifically for time series data, suitable for high-frequency writes, time aggregation, and compressed storage scenarios. It has wide applications in IoT, operations monitoring, financial data, and more. Compared to traditional databases, they have a deeper optimization and understanding of the dimension of “time”, making them one of the essential tools for handling large-scale, continuously changing data.